Detecting tab as CSV separator

[mdruiter]

I often use the Database Viewer for csv files and the like. I always used the "Auto-Select" settings, because those worked okay in most cases.
But version 2.52 changed the rules. And I think they have not been improved.

For 99.9% of the files containing TABs, TAB is the separator. But for many of those files, it now picks some other character.
Example:

Cache.txt

Example

(222 Bytes) Downloaded 255 times

And another one:

2009-11-11.csv

Another example

(109.29 KiB) Downloaded 287 times

Aside: try End in this last example: the long last column overwrites itself...

I now use Tabulator (Tab?) as the default Separator to circumvent the new rule in most cases.

I think the old behavior was "good enough".

[Jan Patera]

For 99.9% of the CSV files containing TABs, TAB is the separator. But for many of those files, DBViewer now picks some other character.

This problem has been fixed for the next release of Salamander (AS 2.53/2.53b1).
Tab as a potential separator now has a higher priority than most other chars and the same as ".:;,". Until now, actually any other character with the _same_ probabilty of being the separator was picked.
Thanks for the sample files.